Google Ads Measurement study materials
Google Ads Measurement study resources

Offline conversion imports that double-count B2B leads

Google Ads Measurement · Put your knowledge to work in context.

All case studies

Offline conversion imports that double-count B2B leads

A B2B SaaS company generates demo requests through a HubSpot form embedded on its site. The marketing ops team imports 'Qualified Lead' and 'Closed Won' offline conversions into Google Ads every night via the API, using the GCLID captured in a hidden field. They also count the website form submission 'Demo Request' as a primary conversion in Google Ads through the GA4 import. The account uses tROAS with a value of $5,000 for Qualified Lead and $40,000 for Closed Won, but values are averages. Since importing offline conversions, reported conversions rose 38%, CPA fell, and Smart Bidding is bidding more on brand terms. Sales says lead quality dropped. The sales cycle is 45-90 days, and only 60% of records have a GCLID. Budget is $50k/month.

1. Diagnose why conversion volume and bidding behavior changed, and identify the signal-quality risks.

Read the worked answer

The rise is likely double counting. A single demo request can later become a qualified lead, so Google Ads counts the form submission and the imported qualified lead as two primary conversions for one lead, and tROAS sees $5,000 of value on top of the form conversion. Smart Bidding then overvalues brand and low-intent form fills. The 38% lift is suspicious because imported offline conversions should overlap with the original lead, not add net-new demand. Only 60% GCLID match means 40% of offline outcomes are invisible, so bidding may overfit to the matched segment, which can differ from the full population. Flat average values make tROAS noisy and can reward the wrong lead types. Brand terms may capture existing demand and inflate apparent return. Diagnose by reconciling CRM lead IDs to GCLIDs, comparing 'All conversions' with 'Conversions,' and segmenting by conversion action and brand vs non-brand.

2. Recommend a conversion action setup and value strategy that fixes counting and handles missing GCLIDs and a long sales cycle.

Read the worked answer

Use one primary conversion action per funnel stage. Make Demo Request primary only if sales accepts form fills; otherwise make it secondary and use Qualified Lead as the sole primary bidding action. Do not bid to both form fill and qualified lead. Import Closed Won as a separate value-reporting action but not as a primary bidding action until volume is sufficient. Use actual deal value or a lead-score-based value rule, not a flat $5,000. For missing GCLIDs, capture GA4 client ID and import enhanced conversions for leads with hashed email or phone; also use offline conversion imports with hashed contact identifiers. Extend the conversion window to 90 days to cover the sales cycle, and use conversion lag reporting to avoid premature optimization. Separate brand into its own campaign with its own budget so brand conversions do not distort non-brand tROAS.

3. What measurement would validate lead quality before scaling budget?

Read the worked answer

Validate with a CRM-matched cohort: compare SQL rate and closed-won rate for Google Ads leads before and after the offline import, matched by GCLID or hashed contact. Use a holdout or spend-matched geo test if possible. Track cost per qualified lead and cost per closed-won, not just CPA. Check whether the bid strategy changed the brand vs non-brand mix; if brand share rose, lead quality may fall. Set a guardrail: if SQL rate drops more than 10% relative, pause value-based bidding and revert to target CPA on qualified leads. Review the match-rate bias by comparing matched vs unmatched leads on industry, company size, and deal value.